13
Key Takeaways
- Cloud computing giants are set to spend a record amount on infrastructure this year, with chipmakers likely to benefit, according to Bank of America analysts.
- Among chip companies, those “highly exposed to data center/AI end markets” could fare better than others, the analysts said.
- One of them is Nvidia, which commands more than 80% of the AI accelerator market, and could maintain its lead over competitors.
Cloud computing giants like Microsoft ( MSFT ) and Amazon ( AMZN ) are set to spend a record amount on infrastructure in 2024, with chipmakers likely to benefit from the race to build out artificial intelligence (AI) data centers, according to Bank of America analysts.
The analysts estimated cloud service provider capital expenditures will hit $121 billion in the second half of 2024, bringing the year’s total to a record $227 billion. That would represent a 39% increase from last year.
Among the largest U.S. providers, Alphabet ( GOOGL ), Microsoft, and Meta ( META ) are expected to spend more than twice as much as they did in 2020, while Oracle ( ORCL ) could boost capex nearly sixfold. The portion of that spending going toward data centers—already about 55%—is expected to continue to rise.
Which Chipmakers Stand To Benefit
Major beneficiaries of that spending are likely to be semiconductor companies like Nvidia ( NVDA ), the analysts said. Nvidia commands a more than 80% share of the AI accelerator market, which BofA forecasts will grow from $100 billion today to $200 billion in 2027. The analysts expect the chipmaker to maintain its dominance throughout the current capex upcycle.
“Overall, we believe those highly exposed to data center/AI end markets should generally fare better,” the analysts wrote. Other top picks include Broadcom ( AVGO ), Micron ( MU ), and Arm ( ARM ).
Intel ( INTC ), meanwhile, is forecast to dramatically trail its peers in terms of exposure to data center spending. With 23% of sales coming from data centers in 2023, it is the chip company analyzed by BofA with the least exposure to AI. That figure is expected to inch up to just 25% in 2026.